[MODEL]

Resolved 💬 3 comments Opened Nov 22, 2025 by aragrow Closed Nov 26, 2025

Preflight Checklist

  • [x] I have searched existing issues for similar behavior reports
  • [x] This report does NOT contain sensitive information (API keys, passwords, etc.)

Type of Behavior Issue

Claude modified files I didn't ask it to modify

What You Asked Claude to Do

I have been experiencing some frustrating moments example I have a document in the .claude directory that clearly identify the difference between Tenant and Client. There are two different entities. And at some point Claude recognized that. Only for Claude later own to totally refactor code and send me the message that there is no difference between Tenant and Client. This cost me time and a ton of credits. I understand that this could happens and I am sure it would be better. I try contacting an actual person to see if there is a way of getting credited for those credit. But no luck. Anybody here has experience the same thing and was able to get a those tokens credited back?

What Claude Actually Did

Claude refactor my code and change the code from tenants to clients.

Expected Behavior

Claude should read the file in the .claude to look at my instructions.

Files Affected

Permission Mode

Accept Edits was ON (auto-accepting changes)

Can You Reproduce This?

Sometimes (intermittent)

Steps to Reproduce

_No response_

Claude Model

Sonnet

Relevant Conversation

Impact

Critical - Data loss or corrupted project

Claude Code Version

2.0.24 (Claude Code)

Platform

Anthropic API

Additional Context

_No response_

View original on GitHub ↗

This issue has 3 comments on GitHub. Read the full discussion on GitHub ↗